In the 1950s, 1960s and early 1970s Britain closed thousands of railway stations. I mapped every closure in Wikipedia's lists, found the towns that lost every station and never got one back, and compared them with similar towns that kept theirs.
On 27 March 1963 the British Railways Board published The Reshaping of British Railways, the report written under its chairman, Richard Beeching. It proposed closing 2,363 stations and about 5,000 miles of line. Many closures came before it and many after, but his name has stuck to all of them.
Ask in almost any small town that lost its trains and someone will tell you the closure was the start of a long decline. Is that true? It is a hard question to answer fairly, because the closures were not random. The lines that went were the quiet ones, and quiet lines served small, rural places. So a town that lost its station might look different today simply because it was always a smaller, sleepier sort of place.
I tried to answer three questions with open data. When did the stations close? Which towns lost their railway for good, and which got it back? And how do those towns compare today with towns of the same size, in the same region, that kept a station?
Wikipedia's List of closed railway stations in Great Britain runs to 12 pages and 6,907 entries, from the first closure in 1833 to the present. I parsed every table and placed 93% of the stations on the map, mostly through their Wikidata entries. For the closures of 1955 to 1975 that share is 96%.
Search for a town, or press play to watch the network thin out year by year. Dots turn bright red in the year they close. Switch to "Towns today" to see which built-up areas lost every station, which kept one and which got one back.
The closures came in two waves. The first, in the early 1950s, came as the Branch Lines Committee, set up in 1949, closed the least-used branch lines. The second, much bigger, peaked in 1964, the year after the report, when 556 of the listed stations closed. Between 1955 and 1975 3,331 stations closed, 48% of all the dated entries in the lists; 1,819 of them closed between 1963 and 1970 alone. By the mid 1970s the closures had all but stopped.
One caution about these counts. Wikipedia keeps separate lists of closed stations for London, Greater Manchester, Merseyside, the West Midlands, Lancashire and Norfolk, which I did not parse, so some urban closures may be missing. I filled in 17 closures of 1955 to 1975 from Wikidata where the lists had no station nearby. Where both sources give a closure year, they agree exactly for 85% of 872 stations.
To turn stations into towns I used the ONS Built-up Areas of 2022, which cover every town and village in Great Britain. A station belongs to a town if it lies within 2 km of the town's centre, or within 1 km of its edge for larger towns (I treat each town as a circle with the same area). Then I sorted every town by its railway history:
For the comparisons I keep to the 2,587 built-up areas in England and Wales with 2,000 to 100,000 people in 2021, because Census 2021 figures for Scotland are not in my data. Of the 1,720 of these towns that had a station in 1955, 643 (37%) lost it and still have none, 78 lost their railway but have it back, and 981 kept a station throughout.
The closures picked small places. Half of the towns of 2,000 to 5,000 people that had a station in 1955 have lost it for good (50%), against 4% of the towns of 50,000 to 100,000. The lost towns are also a little further from big cities: their median distance to a city of 100,000 or more is 18.6 km, against 15.3 km for the towns that kept a station. That is the selection problem in a nutshell, and it is why every comparison below allows for size and region.
The East Midlands and the South West lost the most: 58% and 58% of their station towns. The South East, with its busy commuter lines, lost the fewest, 19%.
Across all of England and Wales, 989 built-up areas lost every station in the closures and still have none; 155 of them have 10,000 people or more, and together they are home to 5,691,918 people. Some are suburbs of cities that are served by trams or metros instead: the Tyne and Wear Metro took over the lines through Wallsend, and Sale has Manchester's Metrolink, which my station data do not count. But many are free-standing towns: Washington (51,878 people, last station closed 1964), Leigh (43,532, 1969), Skelmersdale, Dunstable, Witney, Haverhill, Consett and Portishead, where plans to reopen the line are in hand.
99 built-up areas in England and Wales, and 35 in Scotland, lost every station but have one again. Mansfield, with 61,294 people, was long said to be the largest town in Britain without a railway, until the Robin Hood Line reopened in the 1990s. Corby got its station back in 2009, Kenilworth in 2018, Okehampton's regular trains returned in 2021, and Ashington's in 2024 on the reopened Northumberland Line. In Scotland the Borders Railway brought trains back to Galashiels in 2015.
Now the big question. For each town I took six figures: the share of residents aged 65 or over, the share of households without a car or van, the share of adults in work, the share long-term sick or disabled, the share in managerial or professional jobs (all Census 2021) and mean household income (ONS estimates for the financial year ending 2023, available for 1,198 of the towns).
Raw medians mix up the railway with size: the lost towns are half the size of the kept towns, and small towns are older and richer in cars. So I compared each group with kept towns of similar size in the same region, using a regression with region, size and size squared. The chart shows the gaps and their 95% intervals.
Among towns of the same size today, the differences are modest, but not nothing. Compared with towns of the same size and region that kept a station, the towns that lost theirs have about £1,613 a year lower household income (3.1% of the kept towns' median), and 1.3 percentage points fewer adults in managerial and professional jobs. They have 0.8 points fewer households without a car: life there needs one. On age, work and long-term sickness I find no clear difference at all. The lost towns are not noticeably older (+0.3 points, well within chance).
Size for size, the towns that lost their trains are a little poorer and more car-bound than their neighbours with stations. They did not fall off a cliff.
Where does the income gap come from? A clue: among the 239 lost towns within 15 km of a city of 100,000 or more, the gap is £2,105; among the 404 further away it all but vanishes (£261, easily chance). Near a city a station means a commute, and it would make sense for commuters on good salaries to pick towns with stations. If so, the railway today sorts who lives where more than it decides whether a town lives or dies.
That is not the whole story, and the part my comparison cannot see may be the biggest. Because I compare towns of the same size in 2021, any people a town lost because its trains went are matched away. A careful study by Stephen Gibbons, Stephan Heblich and Edward Pinchbeck, which follows the rail network and local populations from 1950 to 1980 and beyond, finds that places that lost more rail access lost population relative to others, along with jobs, skilled workers and young people. My result is consistent with that: the cuts may have made some towns smaller rather than poorer.
The towns that got their railway back are different again. Many are old industrial towns, such as Corby, Mansfield and Aberdare, and today they have 2.2 points more adults in work than similar kept towns, and fewer older residents. With only 78 of them, and reopenings chosen where the case was strongest, I would not read much into it.
Association, not cause. The closures were chosen. Beeching's report targeted lines and stations with few passengers, which were mostly in smaller, more rural places. I compare towns of similar size in the same region, but towns differ in ways my data do not see: how busy the station was, the local economy in 1960, the roads. And 2021 size may itself reflect the closures, so matching on it could hide some of their effect. Nothing here shows what the closures did; it shows how the towns differ now.
Before and after. I have no town figures from before the closures, so I cannot compare growth. Gibbons, Heblich and Pinchbeck (Journal of Urban Economics, 2024) do exactly that with census data from the 1950s on, and find lasting population losses where rail access was cut.
Which station serves which town. My rule, a station within 2 km of the centre or 1 km of the edge, is rough. Large towns are not circles, so a station on the edge can be missed or one in the next town counted. With a tighter rule (1 km, or 0.5 km beyond the edge) 531 towns count as lost and the income gap is £2,201; with a looser one (3 km, or 1.5 km beyond the edge) 690 towns and £2,152. Leaving out the 58 lost towns with a metro, Underground or heritage station gives £1,566, and keeping only towns of 5,000 or more gives £1,348.
Stations open today. I count National Rail stations only, so trams, metros and heritage railways do not count as keeping a town's railway. I worked out which of today's stations were closed at the end of 1975 by matching them to the closed lists and to Wikidata opening dates; 154 of the 2,609 stations look reopened and 188 new since 1975. This matching will have made some mistakes, particularly where an old station of the same name closed while another stayed open.
Coverage. 473 listed stations could not be placed, and I did not parse Wikipedia's separate lists for London, Greater Manchester, Merseyside, the West Midlands, Lancashire and Norfolk. Income estimates exist only for towns that have a best-fit MSOA of their own, so the smallest towns drop out of the income comparison.
Closed stations. I parsed every table in the 12 pages of Wikipedia's List of closed railway stations in Great Britain (A to Z): station, company, the year in the "Year closed" column (the latest year given, ignoring any reopening year) and any reopening noted. I placed each station by its linked article's Wikidata item (5,916 stations), then by a closed or historic station of the same name in OpenStreetMap (335), then by the place it is named after in OS Open Names when that name is unique in its county (176); 473 remain unplaced. I added 17 Wikidata stations closed in 1955 to 1975 that had no list entry within 500 m.
Stations today. The 2,609 National Rail stations in the uk-railway-stations dataset (October 2026). A station counts as closed at the end of 1975 if a list entry of the same name within 1 km closed between 1950 and 1975 (or earlier, with a reopening noted after 1975), or if Wikidata gives it a reopening date after 1975; as new if Wikidata's earliest opening date is after 1975.
Towns and outcomes. ONS Built-up Areas 2022 for Great Britain; Census 2021 tables TS001, TS007A, TS045, TS062 and TS066 summed from LSOAs by the ONS best-fit lookup; mean household income from the ONS small area estimates (FYE2023), population-weighted over best-fit MSOAs. Comparisons: least-squares regressions of each outcome on the railway group, log population and its square, and region, with robust standard errors; checks add the log distance to the nearest built-up area of 100,000 or more, and a nearest-neighbour match within region on population. These are associations, not causes.